Fix dims and coords returned by compute_forward_vector

[willGraham01]

Description

What is this PR

• Bug fix
• Addition of a new feature
• Other

Why is this PR needed?

See #381

What does this PR do?

The function concerned, compute_forward_vector, has been changed to now explicitly construct the upwards_vector with appropriate labelling, and then the result has the extra spatial dimension dropped to ensure we always drop the correct axis.

The existing unit tests have been given a similar treatment.

• The invalid inputs test has not been touched - this still passes since (functionally) nothing was wrong originally
• The input/output test that was already present has largely stayed the same, save for the addition of additional checks that the coordinates we expect to be preserved, are indeed preserved. The existing numerical computations were valid, so have been left as part of the test.
• The test involving nan values has been reworked to use the more precise sel and isnull method(s) to pull expected / unexpected NaN values out of the result, check that the appropriate time, space, and individuals axes are preserved, AND to force a comparison of the shape of the sub-array that contains NaN values against that which is expected. Previously this test did not actually check if the returned array was empty (which it was), so the np.isnan(...).all() and not np.isnan(...).any() were comparing to the empty container, this always returning True.

References

Closes #381

How has this PR been tested?

Local test suite passes, both prior to and with the additions mentioned above.

In the case of the old nan test, the old version of the test passed with the updates to the code (though this was to be expected due to the aforementioned empty-array bug), before being replaced.

Is this a breaking change?

No

Does this PR require an update to the documentation?

If any features have changed, or have been added. Please explain how the
documentation has been updated.

Checklist:

• The code has been tested locally
• Tests have been added to cover all new functionality
• The documentation has been updated to reflect any changes
• The code has been formatted with pre-commit

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.79%. Comparing base (aac61b0) to head (044e11a).
Report is 2 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #382   +/-   ##
=======================================
  Coverage   99.79%   99.79%           
=======================================
  Files          14       14           
  Lines         969      969           
=======================================
  Hits          967      967           
  Misses          2        2           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[niksirbi]

Thanks a lot for the quick bug fix and the very thorough explanation @willGraham01!

Previously this test did not actually check if the returned array was empty (which it was), so the np.isnan(...).all() and not np.isnan(...).any() were comparing to the empty container, this always returning True.

Thanks for checking this, it makes sense now.

LGTM, barring an optional comment. Feel free to merge at your leisure.
I will make a patch release after this PR is merge.